- Email: [email protected]
Localized corrosion behavior of a zirconium-based bulk metallic glass relative to its crystalline state
Intermetallics 10 (2002) 1157–1162 www.elsevier.com/locate/intermet
Localized corrosion behavior of a zirconium-based bulk metallic glass relative to its crystalline state W.H. Petera, R.A. Buchanana,*, C.T. Liub, P.K. Liawa, M.L. Morrisona, J.A. Hortonb, C.A. Carmichael Jr.b, J.L. Wrightb a
University of Tennessee, Department of Materials Science and Engineering, 434 Dougherty Engineering Building, Knoxville, TN 37996-2200, USA b Oak Ridge National Laboratory, Metals & Ceramics Division, PO Box 2008, MS 6115, Oak Ridge, TN 37831-6115, USA Received 4 June 2002; accepted 15 July 2002
Abstract To date, few detailed corrosion studies of the new bulk metallic glasses (BMGs) have been presented. In the present work, the aqueous electrochemical corrosion properties of BMG-11, 52.5Zr–17.9Cu–14.6Ni–5.0Ti–10.0Al (atomic percent), were investigated. Cyclicanodic-polarization tests were conducted on amorphous and crystalline specimens in a 0.6 M NaCl solution (simulated seawater) and on amorphous specimens in a 0.05 M Na2SO4 solution (simulated moisture condensation, as related to ongoing fatigue experiments in humid air), all at room temperature. In the NaCl solution, both amorphous and crystalline materials were found to exhibit passive behavior with low corrosion rates (15 mm/year or less). However, susceptibilities to pitting corrosion were observed. The amorphous material was found to be more resistant to the onset of pitting corrosion under natural corrosion conditions. In the 0.05 M Na2SO4 solution, the amorphous BMG-11 was found to exhibit passive behavior with a very low corrosion rate (0.4 mm/year), and to be immune to pitting corrosion. Furthermore, when the protective passive film was removed by scratching with a diamond stylus, it was found to quickly reform. This result suggested that a corrosion influence on the fatigue properties of BMG-11 in humid air would be minimal. Published by Elsevier Science Ltd. Keywords: B. Corrosion; B. Glasses, metallic
1. Introduction Recent advances have generated major scientific interest in the field of metallic glasses. Metallic glasses are alloys with no long-range periodic lattice structure [1,2]. Traditionally, fabrication limitations have kept the thicknesses of manufactured specimens often below 50 mm [3,4]. In the early 1990s, a major breakthrough in innovative alloy compositions allowed the production of metallic glasses with thicknesses over one millimeter. Alloys consisting of these new compositions, which can be produced by conventional casting techniques, have been labeled ‘‘Bulk Metallic Glasses’’ (BMGs). Since then, samples with diameters up to several centimeters have been produced [4]. Overall, the corrosion resistances of metallic glasses are expected to be better than those of their crystalline counterparts [5,6]. In metallic glasses, grain boundaries * Corresponding author. Tel.: +1-865-974-4858; fax: +1-865-9744115. E-mail address: [email protected] (R.A. Buchanan). 0966-9795/02/$ - see front matter Published by Elsevier Science Ltd. PII: S0966-9795(02)00130-9
and second-phase precipitates are not present. Ideally, the materials are completely homogenous. This homogeneity can drastically reduce the probability of preferential or localized corrosion. However, one must realize that the basic chemical composition is often a dominating factor in controlling the relative corrosion resistance in all materials, including metallic glasses. Limited corrosion studies of zirconium (Zr)-based metallic glasses have been conducted [6,7]. Hashimoto and Masumoto compared the corrosion rates of a 50Zr– 50Cu (atomic percent, at.%) amorphous alloy (not classified as a ‘‘bulk’’ metallic glass) with those of its crystalline state [6]. In 1 N H2SO4, 1 N HNO3, and 1 N NaOH, the corrosion rates were low and comparable. However, in 3.0 weight percent, (wt.%) NaCl (0.5 M, similar to simulated seawater) and 1 N HCl, the corrosion rates of the amorphous material were approximately one-half those of the crystalline material. Schroeder, Gilbert, and Ritchie conducted corrosion tests on a Zr-based BMG, Zr–13.8Ti–12.5Cu–10Ni– 22.5Be (at.%), in 0.5 M NaCl and compared the results
1158
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
to those of its crystalline form [7]. The amorphous material proved to be slightly more corrosion resistant than the crystalline material, but the results were very similar. Both the amorphous and crystalline materials underwent localized corrosion.
2. Experimental procedures The Zr-based bulk metallic glass investigated in the present study, BMG-11, was originally developed at the California Institute of Technology, Pasadena, California [8,9]. Its chemical composition was 52.5Zr–17.9Cu– 14.6Ni–5.0Ti–10.0Al (at.%). Amorphous round-bar ingots of BMG-11, 6.4 mm in diameter by 76 mm in length, were fabricated at the Oak Ridge National Laboratory (ORNL) using arc-melting and drop-casting procedures. The ingots were cut into metallographic, Xray diffraction, and corrosion samples with an electrical discharge machine. Several of the samples from different ingots were annealed in vacuum at 600 C for 5 h to produce the crystalline state. The metallographic samples were polished to a 0.5 mm diamond-paste finish, etched with a nitric/hydrofluoric acid solution, and examined by optical microscopy. X-ray diffraction analyses were conducted on amorphous and crystalline samples using a Philips X’pert X-Ray Diffractometer. Electrochemical cyclic-anodic-polarization tests were performed on the BMG-11 amorphous and crystalline samples using an EG&G Princeton Applied Research Model 263A Potentiostat/Galvanostat with EG&G 352 SoftCorr III computer software. The electrochemical cell consisted of the corrosion sample, the electrolyte, a saturated calomel reference electrode (SCE), and a platinum counter electrode. The exposed surface of the corrosion sample was a round section transverse to the length dimension of the ingot, with an area of approximately 20 mm2. Before each polarization scan was initiated, the corrosion sample was allowed to stabilize in the electrolyte for either 1 h or until the corrosion potential, Ecorr, changed by no more than 2 mV over a 5-min time period. The scan was started at 50 mV below Ecorr and continued in the positive direction until an anodic current density of 104 mA/m2 was reached. At this point, the scan direction was reversed, and the scan was continued in the negative direction until the original potential was reached. The potential scan rate was 0.17 mV/s. Two electrolytes were used in the investigation: (1) 0.6 M NaCl, which simulated seawater, and (2) 0.05 M Na2SO4, which simulated condensed water vapor and was related to an ongoing fatigue study of BMG-11 in humid air. With regard to the second electrolyte (essentially water), the small amount of Na2SO4, which was regarded as nonaggressive, was added to the distilled water to provide adequate electrical conductivity for the electrochemical experiments.
In the 0.6 M NaCl solution, amorphous and crystalline BMG-11 samples were tested with two surface conditions: (1) ground to a 600-grit SiC finish, which simulated an industrial finish, and (2) metallographically polished to a 0.5 mm diamond-paste finish. For tests in the 0.05 M Na2SO4 solution, the samples were ground to a P2400-grit (Federation of European Producers of Abrasives) SiC finish, which corresponded to the surface finish of the specimens in the ongoing fatigue study [10,11]. Moreover, in one of the polarization tests in the Na2SO4 solution, the sample surface was periodically scratched (at 100 mV intervals for approximately 10 s) with a diamond stylus to intentionally remove the protective passive film in a small region. The anodic current density was monitored to determine whether the film would reform, and if so, the reformation (repassivation) kinetics. This particular test also was related to the fatigue studies, in that, during corrosion fatigue, shear-band movement to the surface will break the passive film. In order for the environment to have a minimal effect on fatigue properties, the passive film must reform quickly. In analyzing the cyclic-anodic-polarization behaviors, a number of corrosion-related parameters were evaluated. These parameters are identified in the schematic polarization curves of Fig. 1, which illustrate typical behaviors with regard to localized corrosion susceptibilities. The plots are potential (relative to the SCE reference electrode) versus the logarithm of the externalcircuit current density, where the current density is the measured external-circuit current divided by the specimen area. The controlled specimen potential can be regarded as the ‘‘driving force’’ for corrosion, and the anodic current density is directly related to the specimen corrosion rate. The potential scan is started below the corrosion potential, Ecorr. At Ecorr, the current density goes to zero, and then increases to a low and approximately constant anodic value ( 10 mA/m2 in Fig. 1) in the passive range. In this range, a thin oxide/hydroxide film, a passive film, protects the material from high
Fig. 1. Schematic cyclic-anodic-polarization curves.
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
corrosion rates. If the current density decreases when the potential scan direction is reversed, as in path 1, the material is shown to be immune to localized corrosion. However, if on the potential upscan, the current density suddenly increases, and remains high on the downscan, until finally decreasing to the passive-region value, as in path 2, the material is shown to undergo a form of localized corrosion (pitting corrosion in this case). The potential at which the current density suddenly increases (pit initiation) is known as the pitting potential, Epit, and the potential at which the current density returns to the passive value is known as the repassivation potential or the protection potential, Epp. Between Epit and Epp, pits are initiating and propagating. In the case of path 2, pits will not initiate at Ecorr, the natural corrosion potential; and, therefore, the material will not undergo pitting corrosion under natural corrosion conditions. If, on the other hand, path 3 is exhibited, where Epp is below Ecorr, the material will undergo pitting corrosion at surface flaws or after incubation time periods at Ecorr. In terms of the overall resistance to pitting corrosion, two parameters are important, (EpitEcorr) and (EppEcorr). Higher values of both are desirable to reflect high values of Epit and Epp relative to Ecorr. The corrosion rate under natural corrosion conditions, i.e. at Ecorr, is related to the corrosion current density, icorr, which must be determined by extrapolative procedures or one of several analytical methods. In this study, the polarization-resistance method was used to evaluate icorr [12]. The corrosion penetration rate (CPR, mm/year) is then calculated by application of Faraday’s law: CPR ¼ 0:327ðMicorr Þ=m
ð1Þ
where M (g/mol), m, and (g/cm3) are the atom-fraction-weighted values of atomic weight, ion valence, and density, respectively, for the alloy elements, and icorr (mA/m2) is the corrosion current density [12].
1159
crystallinity observed in some cross sections of the amorphous, as-cast ingots was in the form of microstarlets forming thin spirals just inside the circumference of the ingot. This feature was consistent with BMG-11 samples previously fabricated and observed at ORNL, and was deemed acceptable for the present study. The crystalline locations were largely covered by an epoxy masking material prior to corrosion testing. The average cyclic-anodic-polarization results for the amorphous and crystalline BMG-11 with 600-grit surface finishes, tested in the 0.6 M NaCl solution, are shown in Fig. 2. Three-to-four replicate tests were run for each condition, with the specimens being taken from different ingots, so that both experimental error and material variability could be evaluated. Both the amorphous and crystalline materials were in the passive state at the natural corrosion potentials, Ecorr. The corrosion current densities (icorr) and corrosion penetration rates (CPRs) were evaluated to be 1.2 0.6 mA/m2 and 1.3 0.7 mm/year, respectively, for the amorphous material, and 0.5 0.3 mA/m2 and 0.6 0.3 mm/year, respectively, for the crystalline material. These corrosion rates are extremely low and reflective of the passive state. Therefore, the concerns for these materials are their resistances to the onset of pitting corrosion. The statistical analyses of the various electrochemical parameters are presented in Fig. 3, where the error bars represent one standard deviation (Epp results exhibited small standard deviations, and the error bars are not visible beyond the symbols used). It is seen that no statistical differences were observed between the amorphous and crystalline materials for the corrosion potentials (Ecorr), pitting potentials (Epit), or (EpitEcorr). However, the differences in protection potentials (Epp) and (EppEcorr) were statistically significantly, with the amorphous material having higher values in both cases. In accordance with prior discussion, the latter results indicate that the amorphous
3. Results and discussion Results of the metallography and X-ray diffraction indicated that the as-cast BMG-11 was an amorphous material, i.e. either no or very little crystallinity was present. The polished and etched specimens showed no structure for the amorphous material and a very-fine grain structure ( < 1 mm grain size) for the crystalline material. The X-ray diffraction patterns revealed a broad peak identifying limited short-range order for the amorphous material, and sharp peaks identifying longrange order for the crystalline material. Relative to the crystalline material, previous studies have shown that the major phases present in BMG-11 after annealing at 600 C are Zr2Cu and Zr2Ni [13,14]. The limited
Fig. 2. Average cyclic-anodic-polarization behaviors of amorphous and crystalline BMG-11 with 600-grit surface finish in an aerated 0.6 M NaCl solution at room temperature.
1160
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
material is more resistant to the onset of pitting corrosion under natural corrosion conditions. The average cyclic-anodic-polarization curves for the amorphous and crystalline BMG-11 specimens in the metallographically-polished condition, tested in the 0.6 M NaCl solution, are shown in Fig. 4. Again, both the amorphous and crystalline materials were in the passive state at the natural corrosion potentials, Ecorr. The corrosion current densities (icorr) and corrosion penetration rates (CPRs) were evaluated to be 8 13 mA/m2 and 9 15 mm/year, respectively, for the amorphous material, and 14 13 mA/m2 and 15 14 mm/year, respectively, for the crystalline material. Although these corrosion rates are approximately 12 times higher, on average, than those evaluated for the rougher 600-grit surface finish, they are still quite low and representative of the passive state. A possible reason for the higher corrosion rates observed for the smoother, polished surface condition (an unusual situation) relates to the time interval between the last surface-finish operation and the immersion of the specimen in the electrolyte.
Fig. 3. Results of statistical analyses of electrochemical parameters for replicate cyclic-anodic-polarization tests of amorphous and crystalline BMG-11 with 600-grit surface finish in an aerated 0.6 M NaCl solution at room temperature.
Fig. 4. Average cyclic-anodic-polarization behaviors of amorphous and crystalline BMG-11 with metallographically-polished surface finish in an aerated 0.6 M NaCl solution at room temperature.
For the polished condition, the time interval was less than 5 min; whereas for the 600-grit condition, it was on the order of 30 min. Therefore, for the polished specimens, less time was available for the formation and thickening of air-formed oxides, and this effect could have resulted in higher corrosion rates. The statistical analyses of the electrochemical parameters for the specimens in the polished condition are given in Fig. 5. It is seen that no statistical differences were observed between the amorphous and crystalline materials for the corrosion potentials (Ecorr) or (EpitEcorr). However, the differences in pitting potentials (Epit), protection potentials (Epp) and (EppEcorr) were statistically significant, with the amorphous material having higher values in all three cases. Again, these latter results indicate that the amorphous material is more resistant than the crystalline material to the onset of pitting corrosion under natural corrosion conditions in the 0.6 M NaCl solution. In the study by Schroeder, Gilbert, and Ritchie [7], anodic polarization tests (not cyclic-anodic-polarization tests) were conducted on a BMG similar to that of this investigation, but containing beryllium (Zr–13.8Ti– 12.5Cu–10Ni–22.5Be, at.%), in a similar electrolyte (0.5 M NaCl), and with a similar metallographicallypolished surface condition (1.0 mm finish). They found that the amorphous state produced higher values of corrosion potential (Ecorr), pitting potential (Epit), and (EpitEcorr) than the crystalline state. These results suggest that the amorphous material is more resistant to pitting corrosion than the crystalline material—a conclusion that would be consistent with the results of the present study on a somewhat different BMG. However, since they did not conduct cyclic-anodic-polarization tests, the protection potentials (Epp) were not evaluated and, therefore, could not be compared to the corrosion potentials, i.e. (EppEcorr). These comparisons yielded important results in the present investigation.
Fig. 5. Results of statistical analyses of electrochemical parameters for replicate cyclic-anodic-polarization tests of amorphous and crystalline BMG-11 with metallographically-polished surface finish in an aerated 0.6 M NaCl solution at room temperature.
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
2.
3.
4. Fig. 6. Anodic-polarization behaviors of amorphous BMG-11 in an aerated 0.05 M Na2SO4 solution at room temperature with the surface undisturbed and with the surface periodically scratched with a diamond stylus.
Cyclic-anodic-polarization curves for BMG-11 in the 0.05 M Na2SO4 solution (essentially water) are presented in Fig. 6. Curves are shown for the undisturbed condition and for a condition where periodic scratching with a diamond stylus was performed. For the undisturbed condition, the amorphous material was in the passive state at the natural corrosion potential, Ecorr, with a corrosion current density (icorr) of 0.4 mA/m2 and a corrosion penetration rate (CPR) of 0.4 mm/year. The corrosion rate was extremely low. On reversing the potential scan direction, the current density decreased in magnitude, indicating that the amorphous material has no susceptibility to localized corrosion, at any potential, in this electrolyte. The periodic scratching (every 100 mV) to locally remove the passive film resulted in drastic anodic-current-density spikes (high local corrosion rates). However, in each case, the current density immediately decreased to the passive value when the scratching was stopped. This behavior indicates the ability for rapid repassivation. As related to corrosionfatigue testing in humid air (where a thin layer of condensed water could exist on the specimen surface), these results indicate that when shear bands move to the surface and create offsets, thereby breaking the oxide film and exposing fresh, unprotected surfaces, those surfaces should immediately passivate. Moreover, a corrosion influence on the fatigue properties of BMG-11 in humid air should be minimal [10,11].
4. Conclusions Based on the results of this investigation, the following conclusions were drawn: 1. Both amorphous and crystalline states of the Zrbased BMG-11 (52.5Zr–17.9Cu–14.6Ni–5.0Ti– 10.0Al, at.%) exhibited passive behaviors at the natural corrosion potentials in the 0.6 M NaCl
5.
1161
solution, with low corrosion penetration rates of 15 mm/year or less. Although both states exhibited pitting-corrosion susceptibilities in the NaCl solution, the electrochemical results indicated that the amorphous material would be more resistant to the onset of pitting corrosion under natural corrosion conditions than the crystalline material. The two surface-finishes, 600-grit and metallographically polished, produced similar electrochemical behaviors in the NaCl solution. In the 0.05 M Na2SO4 solution (simulated condensed water vapor), the amorphous BMG-11 exhibited passive behavior at the natural corrosion potential, with a very low corrosion penetration rate of 0.4 mm/year, and no susceptibility to localized corrosion. When the protective passive film established on amorphous BMG-11 in the dilute Na2SO4 solution was mechanically removed by scratching with a diamond stylus, it quickly reformed. This behavior suggested that a corrosion influence on the fatigue properties of BMG-11 in humid air should be minimal.
Acknowledgements The authors are grateful to the National Science Foundation Integrative Graduate Education and Research Training (IGERT) Program in Materials Lifetime Science and Engineering, managed by Drs. Wyn Jennings and Larry Goldberg, and to the Division of Materials Science and Engineering, Department of Energy under contract DE-AC05–00OR22725 with Oak Ridge National Laboratory (ORNL) operated by UTBattelle, LLC, for support of this research. Also, special thanks are given to Drs. Charlie Brooks, Joseph Spruiell, Roberto Benson, and Charles Feigerle of the University of Tennessee for their advice and consultation.
References [1] Luborsky F. Amorphous metallic alloys. In: Luborsky F, editor. Amorphous metallic alloys. London: Butterworth & Co.; 1983. p. 1. [2] ASM. ASM Handbook. Properties and selections: nonferrous alloys and special purpose materials. Metals Park, OH: ASM; 1990 2:186,667,804. [3] Liu CT, Heatherly L, Easton DS, Carmichael CA, Schneibel JH, Chen CH, et al. Metall Mater Trans A 1998;29A:1811. [4] Inoue A. Intermetallics 2000;8:455. [5] Hashimoto K. Chemical properties. In: Luborsky F, editor. Amorphous metallic alloys. London: Butterworth & Co.; 1983. p. 471.
1162
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
[6] Hashimoto K, Masumoto T. Corrosion properties of amorphous alloys. In: Hasegawa R, editor. Glassy metals: magnetic, chemical, and structural properties. Boca Raton, FL: CRC Press; 1983. p. 241. [7] Schroeder V, Gilbert CJ, Ritchie RO. Scripta Materialia 1998; 38(10):1481. [8] Pecker A, Johnson WL. Appl Phys Letter 1993;63:2242. [9] Lin XH, Johnson WL. J Appl Phys 1995;78:6514. [10] Peter WH, Buchanan RA, Liu CT, Liaw PK. The fatigue behavior of a zirconium-based bulk metallic glass in vacuum and air. Journal of Non-Crystalline Solids (in press).
[11] Peter WH, Liaw PK, Buchanan RA, Liu CT, Brooks CR, Horton Jr. JA, et al. Fatigue behavior of Zr52.5Al10Ti5Cu17.9Ni14.6 bulk metallic glass. Intermetallics (in press). [12] Stansbury EE, Buchanan RA. Fundamentals of electrochemical corrosion. Materials Park, OH: ASM International; 2000. [13] Kundig AA, Lofller JF, Johnson WL. Supercooled liquid, bulk glassy and nanocrystalline states of alloys. In: Inoue A, et al., editors. MRS Sym. Procs Volume 644. Warrendale, PA: MRS Publication; 2001. p. L5.6.1. [14] Nieh TG, Wadsworth J, Liu CT, Ice GE, Chung K-S. Mater Trans, JIM 2001;42:613.
Localized corrosion behavior of a zirconium-based bulk metallic glass relative to its crystalline state W.H. Petera, R.A. Buchanana,*, C.T. Liub, P.K. Liawa, M.L. Morrisona, J.A. Hortonb, C.A. Carmichael Jr.b, J.L. Wrightb a
University of Tennessee, Department of Materials Science and Engineering, 434 Dougherty Engineering Building, Knoxville, TN 37996-2200, USA b Oak Ridge National Laboratory, Metals & Ceramics Division, PO Box 2008, MS 6115, Oak Ridge, TN 37831-6115, USA Received 4 June 2002; accepted 15 July 2002
Abstract To date, few detailed corrosion studies of the new bulk metallic glasses (BMGs) have been presented. In the present work, the aqueous electrochemical corrosion properties of BMG-11, 52.5Zr–17.9Cu–14.6Ni–5.0Ti–10.0Al (atomic percent), were investigated. Cyclicanodic-polarization tests were conducted on amorphous and crystalline specimens in a 0.6 M NaCl solution (simulated seawater) and on amorphous specimens in a 0.05 M Na2SO4 solution (simulated moisture condensation, as related to ongoing fatigue experiments in humid air), all at room temperature. In the NaCl solution, both amorphous and crystalline materials were found to exhibit passive behavior with low corrosion rates (15 mm/year or less). However, susceptibilities to pitting corrosion were observed. The amorphous material was found to be more resistant to the onset of pitting corrosion under natural corrosion conditions. In the 0.05 M Na2SO4 solution, the amorphous BMG-11 was found to exhibit passive behavior with a very low corrosion rate (0.4 mm/year), and to be immune to pitting corrosion. Furthermore, when the protective passive film was removed by scratching with a diamond stylus, it was found to quickly reform. This result suggested that a corrosion influence on the fatigue properties of BMG-11 in humid air would be minimal. Published by Elsevier Science Ltd. Keywords: B. Corrosion; B. Glasses, metallic
1. Introduction Recent advances have generated major scientific interest in the field of metallic glasses. Metallic glasses are alloys with no long-range periodic lattice structure [1,2]. Traditionally, fabrication limitations have kept the thicknesses of manufactured specimens often below 50 mm [3,4]. In the early 1990s, a major breakthrough in innovative alloy compositions allowed the production of metallic glasses with thicknesses over one millimeter. Alloys consisting of these new compositions, which can be produced by conventional casting techniques, have been labeled ‘‘Bulk Metallic Glasses’’ (BMGs). Since then, samples with diameters up to several centimeters have been produced [4]. Overall, the corrosion resistances of metallic glasses are expected to be better than those of their crystalline counterparts [5,6]. In metallic glasses, grain boundaries * Corresponding author. Tel.: +1-865-974-4858; fax: +1-865-9744115. E-mail address: [email protected] (R.A. Buchanan). 0966-9795/02/$ - see front matter Published by Elsevier Science Ltd. PII: S0966-9795(02)00130-9
and second-phase precipitates are not present. Ideally, the materials are completely homogenous. This homogeneity can drastically reduce the probability of preferential or localized corrosion. However, one must realize that the basic chemical composition is often a dominating factor in controlling the relative corrosion resistance in all materials, including metallic glasses. Limited corrosion studies of zirconium (Zr)-based metallic glasses have been conducted [6,7]. Hashimoto and Masumoto compared the corrosion rates of a 50Zr– 50Cu (atomic percent, at.%) amorphous alloy (not classified as a ‘‘bulk’’ metallic glass) with those of its crystalline state [6]. In 1 N H2SO4, 1 N HNO3, and 1 N NaOH, the corrosion rates were low and comparable. However, in 3.0 weight percent, (wt.%) NaCl (0.5 M, similar to simulated seawater) and 1 N HCl, the corrosion rates of the amorphous material were approximately one-half those of the crystalline material. Schroeder, Gilbert, and Ritchie conducted corrosion tests on a Zr-based BMG, Zr–13.8Ti–12.5Cu–10Ni– 22.5Be (at.%), in 0.5 M NaCl and compared the results
1158
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
to those of its crystalline form [7]. The amorphous material proved to be slightly more corrosion resistant than the crystalline material, but the results were very similar. Both the amorphous and crystalline materials underwent localized corrosion.
2. Experimental procedures The Zr-based bulk metallic glass investigated in the present study, BMG-11, was originally developed at the California Institute of Technology, Pasadena, California [8,9]. Its chemical composition was 52.5Zr–17.9Cu– 14.6Ni–5.0Ti–10.0Al (at.%). Amorphous round-bar ingots of BMG-11, 6.4 mm in diameter by 76 mm in length, were fabricated at the Oak Ridge National Laboratory (ORNL) using arc-melting and drop-casting procedures. The ingots were cut into metallographic, Xray diffraction, and corrosion samples with an electrical discharge machine. Several of the samples from different ingots were annealed in vacuum at 600 C for 5 h to produce the crystalline state. The metallographic samples were polished to a 0.5 mm diamond-paste finish, etched with a nitric/hydrofluoric acid solution, and examined by optical microscopy. X-ray diffraction analyses were conducted on amorphous and crystalline samples using a Philips X’pert X-Ray Diffractometer. Electrochemical cyclic-anodic-polarization tests were performed on the BMG-11 amorphous and crystalline samples using an EG&G Princeton Applied Research Model 263A Potentiostat/Galvanostat with EG&G 352 SoftCorr III computer software. The electrochemical cell consisted of the corrosion sample, the electrolyte, a saturated calomel reference electrode (SCE), and a platinum counter electrode. The exposed surface of the corrosion sample was a round section transverse to the length dimension of the ingot, with an area of approximately 20 mm2. Before each polarization scan was initiated, the corrosion sample was allowed to stabilize in the electrolyte for either 1 h or until the corrosion potential, Ecorr, changed by no more than 2 mV over a 5-min time period. The scan was started at 50 mV below Ecorr and continued in the positive direction until an anodic current density of 104 mA/m2 was reached. At this point, the scan direction was reversed, and the scan was continued in the negative direction until the original potential was reached. The potential scan rate was 0.17 mV/s. Two electrolytes were used in the investigation: (1) 0.6 M NaCl, which simulated seawater, and (2) 0.05 M Na2SO4, which simulated condensed water vapor and was related to an ongoing fatigue study of BMG-11 in humid air. With regard to the second electrolyte (essentially water), the small amount of Na2SO4, which was regarded as nonaggressive, was added to the distilled water to provide adequate electrical conductivity for the electrochemical experiments.
In the 0.6 M NaCl solution, amorphous and crystalline BMG-11 samples were tested with two surface conditions: (1) ground to a 600-grit SiC finish, which simulated an industrial finish, and (2) metallographically polished to a 0.5 mm diamond-paste finish. For tests in the 0.05 M Na2SO4 solution, the samples were ground to a P2400-grit (Federation of European Producers of Abrasives) SiC finish, which corresponded to the surface finish of the specimens in the ongoing fatigue study [10,11]. Moreover, in one of the polarization tests in the Na2SO4 solution, the sample surface was periodically scratched (at 100 mV intervals for approximately 10 s) with a diamond stylus to intentionally remove the protective passive film in a small region. The anodic current density was monitored to determine whether the film would reform, and if so, the reformation (repassivation) kinetics. This particular test also was related to the fatigue studies, in that, during corrosion fatigue, shear-band movement to the surface will break the passive film. In order for the environment to have a minimal effect on fatigue properties, the passive film must reform quickly. In analyzing the cyclic-anodic-polarization behaviors, a number of corrosion-related parameters were evaluated. These parameters are identified in the schematic polarization curves of Fig. 1, which illustrate typical behaviors with regard to localized corrosion susceptibilities. The plots are potential (relative to the SCE reference electrode) versus the logarithm of the externalcircuit current density, where the current density is the measured external-circuit current divided by the specimen area. The controlled specimen potential can be regarded as the ‘‘driving force’’ for corrosion, and the anodic current density is directly related to the specimen corrosion rate. The potential scan is started below the corrosion potential, Ecorr. At Ecorr, the current density goes to zero, and then increases to a low and approximately constant anodic value ( 10 mA/m2 in Fig. 1) in the passive range. In this range, a thin oxide/hydroxide film, a passive film, protects the material from high
Fig. 1. Schematic cyclic-anodic-polarization curves.
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
corrosion rates. If the current density decreases when the potential scan direction is reversed, as in path 1, the material is shown to be immune to localized corrosion. However, if on the potential upscan, the current density suddenly increases, and remains high on the downscan, until finally decreasing to the passive-region value, as in path 2, the material is shown to undergo a form of localized corrosion (pitting corrosion in this case). The potential at which the current density suddenly increases (pit initiation) is known as the pitting potential, Epit, and the potential at which the current density returns to the passive value is known as the repassivation potential or the protection potential, Epp. Between Epit and Epp, pits are initiating and propagating. In the case of path 2, pits will not initiate at Ecorr, the natural corrosion potential; and, therefore, the material will not undergo pitting corrosion under natural corrosion conditions. If, on the other hand, path 3 is exhibited, where Epp is below Ecorr, the material will undergo pitting corrosion at surface flaws or after incubation time periods at Ecorr. In terms of the overall resistance to pitting corrosion, two parameters are important, (EpitEcorr) and (EppEcorr). Higher values of both are desirable to reflect high values of Epit and Epp relative to Ecorr. The corrosion rate under natural corrosion conditions, i.e. at Ecorr, is related to the corrosion current density, icorr, which must be determined by extrapolative procedures or one of several analytical methods. In this study, the polarization-resistance method was used to evaluate icorr [12]. The corrosion penetration rate (CPR, mm/year) is then calculated by application of Faraday’s law: CPR ¼ 0:327ðMicorr Þ=m
ð1Þ
where M (g/mol), m, and (g/cm3) are the atom-fraction-weighted values of atomic weight, ion valence, and density, respectively, for the alloy elements, and icorr (mA/m2) is the corrosion current density [12].
1159
crystallinity observed in some cross sections of the amorphous, as-cast ingots was in the form of microstarlets forming thin spirals just inside the circumference of the ingot. This feature was consistent with BMG-11 samples previously fabricated and observed at ORNL, and was deemed acceptable for the present study. The crystalline locations were largely covered by an epoxy masking material prior to corrosion testing. The average cyclic-anodic-polarization results for the amorphous and crystalline BMG-11 with 600-grit surface finishes, tested in the 0.6 M NaCl solution, are shown in Fig. 2. Three-to-four replicate tests were run for each condition, with the specimens being taken from different ingots, so that both experimental error and material variability could be evaluated. Both the amorphous and crystalline materials were in the passive state at the natural corrosion potentials, Ecorr. The corrosion current densities (icorr) and corrosion penetration rates (CPRs) were evaluated to be 1.2 0.6 mA/m2 and 1.3 0.7 mm/year, respectively, for the amorphous material, and 0.5 0.3 mA/m2 and 0.6 0.3 mm/year, respectively, for the crystalline material. These corrosion rates are extremely low and reflective of the passive state. Therefore, the concerns for these materials are their resistances to the onset of pitting corrosion. The statistical analyses of the various electrochemical parameters are presented in Fig. 3, where the error bars represent one standard deviation (Epp results exhibited small standard deviations, and the error bars are not visible beyond the symbols used). It is seen that no statistical differences were observed between the amorphous and crystalline materials for the corrosion potentials (Ecorr), pitting potentials (Epit), or (EpitEcorr). However, the differences in protection potentials (Epp) and (EppEcorr) were statistically significantly, with the amorphous material having higher values in both cases. In accordance with prior discussion, the latter results indicate that the amorphous
3. Results and discussion Results of the metallography and X-ray diffraction indicated that the as-cast BMG-11 was an amorphous material, i.e. either no or very little crystallinity was present. The polished and etched specimens showed no structure for the amorphous material and a very-fine grain structure ( < 1 mm grain size) for the crystalline material. The X-ray diffraction patterns revealed a broad peak identifying limited short-range order for the amorphous material, and sharp peaks identifying longrange order for the crystalline material. Relative to the crystalline material, previous studies have shown that the major phases present in BMG-11 after annealing at 600 C are Zr2Cu and Zr2Ni [13,14]. The limited
Fig. 2. Average cyclic-anodic-polarization behaviors of amorphous and crystalline BMG-11 with 600-grit surface finish in an aerated 0.6 M NaCl solution at room temperature.
1160
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
material is more resistant to the onset of pitting corrosion under natural corrosion conditions. The average cyclic-anodic-polarization curves for the amorphous and crystalline BMG-11 specimens in the metallographically-polished condition, tested in the 0.6 M NaCl solution, are shown in Fig. 4. Again, both the amorphous and crystalline materials were in the passive state at the natural corrosion potentials, Ecorr. The corrosion current densities (icorr) and corrosion penetration rates (CPRs) were evaluated to be 8 13 mA/m2 and 9 15 mm/year, respectively, for the amorphous material, and 14 13 mA/m2 and 15 14 mm/year, respectively, for the crystalline material. Although these corrosion rates are approximately 12 times higher, on average, than those evaluated for the rougher 600-grit surface finish, they are still quite low and representative of the passive state. A possible reason for the higher corrosion rates observed for the smoother, polished surface condition (an unusual situation) relates to the time interval between the last surface-finish operation and the immersion of the specimen in the electrolyte.
Fig. 3. Results of statistical analyses of electrochemical parameters for replicate cyclic-anodic-polarization tests of amorphous and crystalline BMG-11 with 600-grit surface finish in an aerated 0.6 M NaCl solution at room temperature.
Fig. 4. Average cyclic-anodic-polarization behaviors of amorphous and crystalline BMG-11 with metallographically-polished surface finish in an aerated 0.6 M NaCl solution at room temperature.
For the polished condition, the time interval was less than 5 min; whereas for the 600-grit condition, it was on the order of 30 min. Therefore, for the polished specimens, less time was available for the formation and thickening of air-formed oxides, and this effect could have resulted in higher corrosion rates. The statistical analyses of the electrochemical parameters for the specimens in the polished condition are given in Fig. 5. It is seen that no statistical differences were observed between the amorphous and crystalline materials for the corrosion potentials (Ecorr) or (EpitEcorr). However, the differences in pitting potentials (Epit), protection potentials (Epp) and (EppEcorr) were statistically significant, with the amorphous material having higher values in all three cases. Again, these latter results indicate that the amorphous material is more resistant than the crystalline material to the onset of pitting corrosion under natural corrosion conditions in the 0.6 M NaCl solution. In the study by Schroeder, Gilbert, and Ritchie [7], anodic polarization tests (not cyclic-anodic-polarization tests) were conducted on a BMG similar to that of this investigation, but containing beryllium (Zr–13.8Ti– 12.5Cu–10Ni–22.5Be, at.%), in a similar electrolyte (0.5 M NaCl), and with a similar metallographicallypolished surface condition (1.0 mm finish). They found that the amorphous state produced higher values of corrosion potential (Ecorr), pitting potential (Epit), and (EpitEcorr) than the crystalline state. These results suggest that the amorphous material is more resistant to pitting corrosion than the crystalline material—a conclusion that would be consistent with the results of the present study on a somewhat different BMG. However, since they did not conduct cyclic-anodic-polarization tests, the protection potentials (Epp) were not evaluated and, therefore, could not be compared to the corrosion potentials, i.e. (EppEcorr). These comparisons yielded important results in the present investigation.
Fig. 5. Results of statistical analyses of electrochemical parameters for replicate cyclic-anodic-polarization tests of amorphous and crystalline BMG-11 with metallographically-polished surface finish in an aerated 0.6 M NaCl solution at room temperature.
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
2.
3.
4. Fig. 6. Anodic-polarization behaviors of amorphous BMG-11 in an aerated 0.05 M Na2SO4 solution at room temperature with the surface undisturbed and with the surface periodically scratched with a diamond stylus.
Cyclic-anodic-polarization curves for BMG-11 in the 0.05 M Na2SO4 solution (essentially water) are presented in Fig. 6. Curves are shown for the undisturbed condition and for a condition where periodic scratching with a diamond stylus was performed. For the undisturbed condition, the amorphous material was in the passive state at the natural corrosion potential, Ecorr, with a corrosion current density (icorr) of 0.4 mA/m2 and a corrosion penetration rate (CPR) of 0.4 mm/year. The corrosion rate was extremely low. On reversing the potential scan direction, the current density decreased in magnitude, indicating that the amorphous material has no susceptibility to localized corrosion, at any potential, in this electrolyte. The periodic scratching (every 100 mV) to locally remove the passive film resulted in drastic anodic-current-density spikes (high local corrosion rates). However, in each case, the current density immediately decreased to the passive value when the scratching was stopped. This behavior indicates the ability for rapid repassivation. As related to corrosionfatigue testing in humid air (where a thin layer of condensed water could exist on the specimen surface), these results indicate that when shear bands move to the surface and create offsets, thereby breaking the oxide film and exposing fresh, unprotected surfaces, those surfaces should immediately passivate. Moreover, a corrosion influence on the fatigue properties of BMG-11 in humid air should be minimal [10,11].
4. Conclusions Based on the results of this investigation, the following conclusions were drawn: 1. Both amorphous and crystalline states of the Zrbased BMG-11 (52.5Zr–17.9Cu–14.6Ni–5.0Ti– 10.0Al, at.%) exhibited passive behaviors at the natural corrosion potentials in the 0.6 M NaCl
5.
1161
solution, with low corrosion penetration rates of 15 mm/year or less. Although both states exhibited pitting-corrosion susceptibilities in the NaCl solution, the electrochemical results indicated that the amorphous material would be more resistant to the onset of pitting corrosion under natural corrosion conditions than the crystalline material. The two surface-finishes, 600-grit and metallographically polished, produced similar electrochemical behaviors in the NaCl solution. In the 0.05 M Na2SO4 solution (simulated condensed water vapor), the amorphous BMG-11 exhibited passive behavior at the natural corrosion potential, with a very low corrosion penetration rate of 0.4 mm/year, and no susceptibility to localized corrosion. When the protective passive film established on amorphous BMG-11 in the dilute Na2SO4 solution was mechanically removed by scratching with a diamond stylus, it quickly reformed. This behavior suggested that a corrosion influence on the fatigue properties of BMG-11 in humid air should be minimal.
Acknowledgements The authors are grateful to the National Science Foundation Integrative Graduate Education and Research Training (IGERT) Program in Materials Lifetime Science and Engineering, managed by Drs. Wyn Jennings and Larry Goldberg, and to the Division of Materials Science and Engineering, Department of Energy under contract DE-AC05–00OR22725 with Oak Ridge National Laboratory (ORNL) operated by UTBattelle, LLC, for support of this research. Also, special thanks are given to Drs. Charlie Brooks, Joseph Spruiell, Roberto Benson, and Charles Feigerle of the University of Tennessee for their advice and consultation.
References [1] Luborsky F. Amorphous metallic alloys. In: Luborsky F, editor. Amorphous metallic alloys. London: Butterworth & Co.; 1983. p. 1. [2] ASM. ASM Handbook. Properties and selections: nonferrous alloys and special purpose materials. Metals Park, OH: ASM; 1990 2:186,667,804. [3] Liu CT, Heatherly L, Easton DS, Carmichael CA, Schneibel JH, Chen CH, et al. Metall Mater Trans A 1998;29A:1811. [4] Inoue A. Intermetallics 2000;8:455. [5] Hashimoto K. Chemical properties. In: Luborsky F, editor. Amorphous metallic alloys. London: Butterworth & Co.; 1983. p. 471.
1162
W.H. Peter et al. / Intermetallics 10 (2002) 1157–1162
[6] Hashimoto K, Masumoto T. Corrosion properties of amorphous alloys. In: Hasegawa R, editor. Glassy metals: magnetic, chemical, and structural properties. Boca Raton, FL: CRC Press; 1983. p. 241. [7] Schroeder V, Gilbert CJ, Ritchie RO. Scripta Materialia 1998; 38(10):1481. [8] Pecker A, Johnson WL. Appl Phys Letter 1993;63:2242. [9] Lin XH, Johnson WL. J Appl Phys 1995;78:6514. [10] Peter WH, Buchanan RA, Liu CT, Liaw PK. The fatigue behavior of a zirconium-based bulk metallic glass in vacuum and air. Journal of Non-Crystalline Solids (in press).
[11] Peter WH, Liaw PK, Buchanan RA, Liu CT, Brooks CR, Horton Jr. JA, et al. Fatigue behavior of Zr52.5Al10Ti5Cu17.9Ni14.6 bulk metallic glass. Intermetallics (in press). [12] Stansbury EE, Buchanan RA. Fundamentals of electrochemical corrosion. Materials Park, OH: ASM International; 2000. [13] Kundig AA, Lofller JF, Johnson WL. Supercooled liquid, bulk glassy and nanocrystalline states of alloys. In: Inoue A, et al., editors. MRS Sym. Procs Volume 644. Warrendale, PA: MRS Publication; 2001. p. L5.6.1. [14] Nieh TG, Wadsworth J, Liu CT, Ice GE, Chung K-S. Mater Trans, JIM 2001;42:613.